Ph.D. Physics, Boston University (2018)
Sc.B., Physics, Massachusetts Institute of Technology (2012)
My primary research uses data from proton-proton collisions at the CERN Large Hadron Collider (LHC) to probe our understanding of the Standard Model of particle physics. This consists of both direct searches for new physics as well as precision measurements of known particles, particularly in the highly boosted regime. As part of this research I make use of machine learning (ML) for an array of tasks, including jet classification, mass regression, and event reconstruction. I see ML as an important tool for making the most of the enormous volume of data we have collected (and will collect) at the LHC, especially as we begin to study more complex processes with subtle signatures.
My research is also focused on trigger and data acquisition systems at the LHC. The LHC collides protons at a rate of 40 million events per second, but the information from each event cannot be stored, due both to bandwidth and overall storage limitations. Instead, the trigger system must analyze each event rapidly and determine if the event should be saved or discarded forever. This is done in multiple stages, on hardware from custom electronics to field programmable gate arrays (FPGAs) to commercial CPUs to GPUs. My research aims to understand how best to utilize this hardware to maximize the physics that can be done. I see ML as an extremely promising technique for making efficient use of information in latency/throughput-constrained systems. I am a leading member of the FastML collaboration [https://fastmachinelearning.org/], which focuses on problems in this space and supports the hls4ml tool for low-latency ML inference on FPGAs. I am also generally interested in the use of ML across scientific disciplines, particularly in similar cases where model size and latency are constrained.
- CMS Collaboration, “Inclusive search for highly boosted Higgs bosons decaying to bottom quark-antiquark pairs in proton-proton collisions at sqrt(s) = 13 TeV”, J. High Energy Phys. 12, 85 (2020), doi:10.1007/JHEP12(2020)085, arXiv:2006.13251.
- A. Gunny, D. Rankin et al. “Hardware-accelerated inference for real-time gravitational-wave astronomy.” Nat Astron (2022), doi:10.1038/s41550-022-01651-w
- S. E. Park et al. "Quasi anomalous knowledge: searching for new physics with embedded knowledge." Journal of High Energy Physics 2021.6 (2021): 1-26.
- D. Rankin et al. "FPGAs-as-a-service toolkit (FaaST)." 2020 IEEE/ACM International Workshop on Heterogeneous High-performance Reconfigurable Computing (H2RC). IEEE, 2020.
- J. Krupa et al. "GPU coprocessors as a service for deep learning inference in high energy physics." Machine Learning: Science and Technology 2.3 (2021): 035005.
- J. Duarte et al. "FPGA-accelerated machine learning inference as a service for particle physics computing." Computing and Software for Big Science 3.1 (2019): 1-15.
- N. Tarafdar et al. "AIgean: An Open Framework for Machine Learning on Heterogeneous Clusters," 2020 IEEE 28th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), (2020), pp. 239-239, doi: 10.1109/FCCM48280.2020.00072.
- P. Harris et al. ”Physics Community Needs, Tools, and Resources for Machine Learning”, arXiv preprint arXiv:2203.16255 (2022)
- E. Khoda, D. Rankin, R. Teixeira de Lima et al. “Ultra-low latency recurrent neural network inference on FPGAs for physics applications with hls4ml”, Submitted to Mach. Learn.: Sci. Technol. (2022) arXiv:2207.00559
- A. Gunny, D. Rankin et al. “A Software Ecosystem for Deploying Deep Learning in Gravitational Wave Physics”, FlexScience '22: Proceedings of the 12th Workshop on AI and Scientific Computing at Scale using Flexible Computing Infrastructures, (2022), doi: 10.1145/3526058.3535454